Which of these semis would you buy?

Upgrading Taiwan Semi?

I received this note from an investment bank upgrading Taiwan
Semiconductor early this week: The basis was valuation — stock trading at 12 x
2006 EPS, “mid-cycle” valuation; so this analyst believes the price is right.

The analyst had been neutral on
Quote |
Chart |
News |
on valuation and
aggressive capacity adds relative to
Quote |
Chart |
News |
, especially in 300mm in a muted
pricing environment, excess 200mm capacity and overall risk to 2006 estimates
(italics mine).

The bank also rates UMC a buy. According to the analyst, UMCs
foundry has underperformed the fabless companies by 15%. So, the analyst expects
a rebound in stocks as capacity tightens and TSM, UMC take the lead in advanced
technology production (95% share of global 300mm logic capacity, 63% including
Sony versus 75% in 200mm cycle). Contrary to fears of increasing competition
among foundries with the second tier companies (Semtech and Infineon) and new
entrants, this analyst believes UMC and TSM will be the power-houses. The risk
of 1Q inventory adjustment looks priced into the stocks and the analyst is
estimating -10% sequentially in 1Q and sub seasonal quarterly growth through
2006 to show 15% revenue growth. TSM tends to gain share in periods of inventory
adjustments. Capacity growth will slow: first half 2005, 5%; second half 2005,
9% to first half 2006, 2%.”

How is it TSM (and UMC) will be able to increase prices while
SMIC (that big mega-fab HQ’ed in Shanghai) sits with excess capacity? (SMIC has
plenty of capacity coming on-stream plus current capacity building DRAMs is just
“filler”; it would just love to raise prices just a little to capture “logic”
wafers from anybody and everybody…) It seems to me any price increases,
especially in this environment where there is a ton of “me too” competition and
it is difficult to pass them through, will drive prospective customers to SMIC,
et al.

In fairness, the analyst seems to be arguing that just a
utilization increase is enough to drive the bottom line, which is at least
partly true. However, in the past examples the analyst is citing, they were able
to raise prices as capacity tightened. It seems raising prices will be much more
difficult this time around. In turn, it seems likely their bottom-line
performance improvement will be more difficult as well.

Melanie Hollands


For 14 years, Melanie Hollands has covered the
technology and telecommunications sectors, from positions held in
business strategy (McKinsey & Co., Bain & Co.), corporate finance (Salomon Smith
Barney) and fundamental equity research (Merrill Lynch). She follows
PC/server/storage hardware, enterprise and application software, wireless
hardware/software/middleware, data networking and telecom equipment, optics,
semiconductors, semi capital equipment, and various niche technologies (RFID,
WiMax, VOIP and others). Hollands is president of Koala Capital (located in
Aspen, Colo., and New York City), which focuses on trading/investing in
technology stocks. She is also a senior advisory board member for a start-up
financial services venture, the Semiconductor Futures Exchange Inc., and an
advisory board member for a start-up Linux-HPC venture, Tadpole Ventures, LLC.
She has been a guest lecturer and adjunct professor at Columbia Business School,
where she earned her MBA, and serves as an advisory member for various faculty
departments. She also holds a joint bachelor’s in Architecture and Structural